Avian vocal learning is chosen for intensive study for two reasons. On the one hand, it shares many basic characteristics with the process of speech development in man. Experimentation of a kind which could not be conducted with children, as for example on the role in learning of the organism's capacity to hear its own voice in normal ontogeny, is expected to illuminate analogous processes underlying speech development in children. In addition, birdsong learning illustrates in particularly sharp relief the occurrence of constraints on learning ability, as manifest under natural conditions. Several species will be used as subjects selected for study of the role of social predispositions in guiding song learning in certain directions. We propose a series of experiments to identify the nature and timing of events occurring in song development that eventually make the bird refractory to further modification of vocal behavior through learning. The role of auditory pre-expectations in development of genetically-based song differences will be explored. The issues raised by this research, seen in sharp relief in birdsong development, are germane to other types of learning, where their relevance has yet to be adequately acknowledged.